KeyDB is an open-source, in-memory data structure store, designed to act as a database, cache, and message broker. It is a high-performance fork of Redis, meaning it supports the same data structures (strings, hashes, lists, sets, sorted sets, bitmaps, hyperloglogs, geospatial indexes, and streams) and is fully compatible with the Redis protocol, changes, and modules.
The community is smaller than Redis. While the core engine is robust, you will rely on EQ Alpha’s corporate backing rather than the massive Redis OSS community.
Integrates RocksDB directly into the KeyDB core.
This public link is valid for 7 days and shares a thread, including any personal information you added. This link or copies made by others cannot be deleted. If you share with third parties, their policies apply. Can’t copy the link right now. Try again later. KeyDB - The Faster Redis Alternative keydb eng
# /etc/keydb/keydb.conf # Specify the number of threads for the KeyDB engine # Ideally matched to the number of available physical CPU cores server-threads 4 # Enable active-active replication if configuring multiple masters active-replica yes Use code with caution. To run KeyDB via Docker, use the official image: docker run -d --name my-keydb -p 6379:6379 eqalpha/keydb Use code with caution. Ideal Use Cases
Because KeyDB is fully compatible with Redis, your existing Redis clients (e.g., Jedis, Node-Redis, Lettuce) will work without changes. Conclusion
| Metric | Standard Redis (Single Thread) | KeyDB (Multithreaded) | | :--- | :--- | :--- | | | Limited by single-core speed. | Scales linearly with cores. | | Latency (p99) | Spikes during heavy loads due to thread blocking. | Remains stable; background tasks offloaded. | | CPU Utilization | 100% on one core, idle on others. | Even distribution across all cores. | KeyDB is an open-source, in-memory data structure store,
Independent benchmarks have consistently shown KeyDB outperforming Redis in throughput and latency, especially under high concurrency. However, performance depends heavily on workload, hardware, and configuration.
Redis uses a single main thread to process all commands. While this simplifies internals (no locks, predictable performance), it leaves modern multi-core CPUs underutilized.
Using smaller, fewer servers to achieve the same performance as a larger Redis cluster. While the core engine is robust, you will
Traditional Redis relies on a single-threaded event loop to execute data store commands. While newer Redis versions offload network I/O to secondary threads, the execution of commands remains strictly sequential on a single CPU core.
Selecting and tuning the right KeyDB engine allows you to balance ultra-low latency, memory costs, and data persistence. 1. The Core Engines: In-Memory vs. Flash
: An MVCC implementation allows for non-blocking queries like SCAN and KEYS to run without degrading the performance of active workloads.
With Snap Inc. now backing the project, KeyDB has a permanent home and a strong commitment to open source. The entire codebase – including features that were previously part of a commercial offering – is now available under the BSD‑3 license.